Amazon: Retail continues to improve efficiency, huge potential in AI and cloud computing space.

Here is a summary of Amazon's Q4 2023 earnings conference call. For an interpretation of the earnings report, please refer to " "The Resurgence of Amazon: How Many Surprises Are There?""

I. Review of Key Financial Highlights:

1. Steady Growth and Surpassing Profit Expectations: In terms of overall performance, Amazon achieved a total revenue of $170 billion this quarter, nearly $4 billion higher than market expectations. The revenue growth rate continued to improve MoM by 1.3% to 13.9%, mainly driven by the improvement in the growth rate of international business revenue due to the depreciation of the US dollar and the strong performance during the North American Q4 promotion period.

The overall operating profit reached $13.2 billion, approximately $2.8 billion higher than market expectations, or 26%. This is the biggest highlight of this earnings report. The significant improvement in the profit margin of the North American retail business exceeded expectations.

2. Stable Revenue and Significant Profit Increase for Amazon: This quarter, Amazon's revenue reached $24.2 billion, with a MoM growth rate slightly increasing to about 13%, which is in line with market expectations. The market consensus is that Amazon's growth rate will marginally rebound due to the optimization cycle of IT spending by European and American enterprises and the incremental demand brought by AI investment.

However, compared with the significant acceleration of Azure and GCS in the past two quarters, Amazon is currently only stabilizing at the bottom. This still reflects Amazon's relative competitive disadvantage in AI capabilities, which is the main concern in the market.

In terms of profit, Amazon achieved an operating profit of $7.2 billion this quarter, with an operating profit margin slightly decreasing from over 30% to 29.6% MoM, far exceeding the market expectation of $6.6 billion. Although the growth is only a bottoming rebound in line with expectations, the improvement in profit is significantly stronger. The decrease in R&D expense ratio is a visible reason, and it remains to be seen if there are other explanations from the management.

3. Exchange Rate Reversal and Strong Q4 Promotion Drive Retail Segment Growth: This quarter, the company's retail segment achieved a revenue of $145.8 billion, with a YoY growth rate of 14%, a further increase of 1.4% compared to the previous quarter.

On the one hand, the improvement is mainly due to the positive impact of the depreciation of the US dollar, with the growth rate of US dollar revenue in international regions reaching 17% (13% excluding the impact of exchange rates). The growth rate in the North American region, unaffected by exchange rates, also increased by 2% to 13%, showing clear signs of recovery. Recent reports attribute this to strong consumer spending during the Thanksgiving and Christmas holiday promotions.

**In terms of business segments, advertising revenue growth remains strong, with a growth rate of 26.8%. Dolphin Research believes that this is partly due to the continuation of the trend of advertising investment shifting towards performance-based advertising. Additionally, the company recently announced its business of providing TV advertising on Prime Video streaming media, which is expected to continue driving the growth of advertising revenue.4. Profitability of the retail sector reaches a historical high: The most surprising aspect of this performance is the significant improvement in the profitability of the North American retail business. The operating profit of the North American retail sector this quarter reached nearly 6.5 billion, exceeding the expected 4.2 billion by about 52%. The operating profit margin also increased to 6.1% MoM, the second-highest in the history of the North American business. Dolphin Research believes that this is due to the continuous increase in high-profit advertising revenue and the reduction in investment in content service providers, as well as the optimization of unit costs resulting from the decentralization of the North American fulfillment area.

Considering the expenditure, the fulfillment expense ratio has decreased slightly by 0.1-0.2 percentage points YoY and MoM. Although the magnitude is not significant, it reflects the improvement in fulfillment efficiency. We believe that the benefits of decentralization will be fully manifested in the subsequent quarters. On the other hand, the R&D and content expense ratios have decreased significantly by 1 percentage point YoY and 1.8 percentage points MoM, which is a major contribution to the improvement in profitability. Dolphin Research believes that this reflects the company's downsizing in R&D and online audiovisual business, as well as the reduction in investment. It is also one of the main reasons for the significant improvement in the profitability of the retail sector.

1. Positive guidance continues: For the first quarter performance guidance, the company expects the midpoint of revenue to be 140.8 billion, which may seem lower than the market's expectation of 142 billion. However, considering that the company's actual deliveries generally align with the upper limit of the guidance range, the upper limit of 143.5 billion is still higher than the market's expectation. The operating profit guidance range is 8-12 billion, with the midpoint also lower than the expected 9.1 billion, but still nearly 1/3 higher than the expectation when considering the upper limit of the guidance. It can be seen that the company's guidance is relatively optimistic.

2. Overview as follows:

II. Detailed content of the earnings conference call

2.1. Key points from executive statements:

1. Capital expenditure in 2024 is expected to increase YoY, mainly driven by infrastructure spending to support investments in Amazon, generative AI, and large models. Infrastructure capital expenditure accounted for 60% of total expenditure in the company's core business in 2023.

2. Starting from January 2024, the lifespan of servers will be extended from 5 years to 6 years. This will increase revenue by approximately 900 million in 1Q24.

3. Launch of Amazon Q Code Assistant, highlighting that Q is a "true code expert"; launch of Rufus Expert Shopping Assistant, seamlessly integrated with the Amazon shopping website, allowing customers to view summary reviews.

4. Capital expenditure in 2024 is expected to increase YoY, mainly due to increased infrastructure capital expenditure to support the growth of Amazon's business, including additional investments in AIGC and LLM.

5. The rapid growth of daily necessities business. Our regionalization efforts have also shortened transportation distances, which helps reduce service costs. In 2023, for the first time since 2018, we have reduced unit service costs globally. In the United States alone, the unit service cost has decreased by more than $0.45 compared to the previous year.

6. We have increased sponsored TV advertisements in the United States, where brands can create streaming TV ads without any minimum spending requirements. Through Twitch, Fire TV, and the recently launched Prime Video programs and movies in the United States, we are reaching audiences who are actively engaged.

7. Through the inaugural "Black Friday" football game, we have introduced a new NFL tradition, with a 24% increase in viewership and a significant YoY growth in advertising revenue. We are increasingly confident that Prime Video itself can become a large-scale and profitable business.

2.2 Q&A Analyst Questions

Q: Can you talk about the contribution of backlog conversion and AI workloads, as well as the factors that will accelerate Amazon's revenue in the fourth quarter, and how we should consider these factors in terms of the pace of exiting in 2024? And based on your extensive comments on capital expenditures, how should we think about AI-driven capital expenditures in Amazon's plans, as well as broader capital expenditure comments? Thank you.

As of December 31, Amazon's backlog balance is $155.7 billion, an increase of over $45 billion YoY and over $20 billion QoQ.

Revenue growth in the fourth quarter accelerated to 13.2%. This is an acceleration, and we expect the trend to continue until 2024. We are also excited about the interest in generative AI and products like Bedrock and Q, as described by Andy.

Capital expenditures in 2023 were $48 billion, a decrease of $10 billion YoY. Many of the investments in 2023 are related to infrastructure, with infrastructure supporting both Amazon and its core business accounting for about 60% of our expenditures. We expect these trends to continue until 2024, with an increase in capital expenditures in 2024.

I won't provide a specific number today. We are still finalizing this year's plan, but we do expect capital expenditures to increase as Amazon expands. This is primarily due to the work we are doing in generative AI projects in fulfillment centers and logistics.

Firstly, I believe that most of the cost optimization has already been achieved. This is not to say that there won't be further cost optimization or that we don't see opportunities for more cost optimization, but the magnitude of cost optimization has significantly decreased. At the same time, **migration work (which is related to some backlog work) has started to recover, but the migration pace may not be as fast as before.We have also seen that some transactions that usually sign contracts quickly but sign contracts slowly in uncertain environments have been completed in the previous quarter. As for Gen AI, if you look at our Gen AI revenue, in absolute terms, it is still relatively small. We truly believe that we will drive billions of dollars in revenue in the coming years, but what is encouraging is that its growth rate is so fast.

Q: The first question is about the comment on service costs, which has seen its first decline since 2018. As you look ahead to 2024 and 2025, what are the key operational disruptions and response measures that must be taken in order to continue reducing service costs to the level of 2018, or from whichever perspective you consider "Polaris"? The second question is about the concept of capital return. It seems that cash balance can start to accumulate well. How do you view the idea of buybacks, stock repurchases, or some form of capital return plan to help shareholders?

A: Regarding the decline in service costs, I don't think we feel that we will eventually reach that level (2018), but there is still a lot of room for improvement in this regard. As we scale, we are doing better in controlling fixed costs. This is also one of the reasons why we are able to lower the cost of each service.

This is not only reflected in operations, but actually throughout the entire company. We have seen that in 2021 and 2022, some inflationary factors have had a particularly severe impact on us, such as transportation or shipping services, fuel, and others. Therefore, although we have not completely overcome the challenges, the inflation rate is decreasing and we still see more room for improvement.

Regarding the question about stock repurchases. We have gone through a turbulent period, as Andy mentioned, our logistics footprint has doubled and we have made a lot of investments, and you can see that free cash flow is negative, at least in our calculations, free cash flow was negative from 2021 to 2022. Therefore, we are pleased to see the improvement and rebound in free cash flow, and we discuss and discuss capital structure policies every year or more frequently. Today, I have nothing to announce, but we reiterate that we have many strong investments ahead of us.

Q: In the past few quarters, you have seen a good improvement in the profitability of international business. Can you talk about your considerations for increasing operating income to positive numbers and some levers that may bring international business closer to the level of North America over time? Also, do you currently see any shipping disruptions related to the Red Sea? Will this affect your outlook for the first quarter? Thank you.

A: Let me start with the second question. We have noticed global geopolitical issues, especially the supply chain issues you mentioned, and the impact they may have on freight in the United States and Europe. This will not have a substantial impact on our first-quarter guidance. But as I said, we remain vigilant and will make efforts to take necessary measures to ensure that the customer experience is not affected.Regarding the revenue of our international business department, we are very satisfied with the performance, especially in the past few quarters. Our revenue has increased by $1.8 billion YoY. I attribute this to the steady progress we have made in the United States, as mentioned by Andy, including cost reduction in services, increased advertising efforts, emphasis on cost and investment, and control over fixed costs. As a result, we have also seen this in mature countries such as Europe and Japan.

I would like to divide this section into several categories. First, the mature markets in Europe. If you look at emerging countries, we have launched in 10 countries in the past seven years, and they are all on their respective profitable and customer development trajectories, which we are very satisfied with.

In the fourth quarter, our spending on digital content was a bit high due to the many marketing and content initiatives, especially around sports broadcasting, such as the English Premier League and UEFA Champions League in Germany and Italy. But we like these revenues and investments. As I mentioned, they are an effective customer acquisition tool that allows people to shop on our website and enjoy discounts, which always has a positive impact on our relationship with Amazon.

Q: The profit margin in the North American region has improved for seven consecutive quarters or similar quarters, which is a significant figure. I believe that factors such as improving capacity utilization, enhancing regional center efficiency, mitigating shipping, transportation, and logistics costs, labor costs, etc., are likely to continue to contribute to the improvement of our profit margin in North America. If you have any comments on this, please share.

Secondly, regarding prime time videos and Amazon Prime Video, could you provide some background information on the expectations in this area? You have a large number of Prime users, your CPM is reasonable, and the ad load is not high, but it seems that you should have a great opportunity. Can you explain the scale or how you view the growth potential?

A: We have already provided a good explanation for the previous question regarding cost structure, regionalization, growth of new assets during the pandemic, high efficiency and productivity of all operational networks, focus on fixed costs, and reducing costs, maintaining costs, and reducing advertising. We have successfully achieved advertising revenue growth that exceeds the growth rate of traffic. We expect all these trends to continue, and we have a guiding indicator, perhaps the profitability before the pandemic, but we will not limit ourselves to our improvements. We will continue to look for ways to reduce service costs. At the same time, we may increase customer experience.

Regarding the second question about advertising, I cannot answer it at the moment. For video advertising, advertisers are pleased to reach our main customer base. We are working on increasing our advertising and streaming business, including Fire TV, Prime Video, Freebie, and Twitch, which are important components of the entire business model. We hope it will give us a healthy business, continue to invest in content, and continue to grow. We feel good about it.Q: First, let's talk about groceries. Can you please discuss the progress made in terms of products between Fresh and Whole Foods? How can we reduce the cost of reverse logistics and are there significant opportunities in driving traffic and revenue in the grocery business?

Next is healthcare. It is an industry with a poor customer experience and a notorious reputation. You have made significant efforts in acquisitions and providing primary care. Can you further discuss the long-term vision for the healthcare industry?

A: In terms of groceries, we are pleased with the progress we have made. We have been considering our current grocery business, which I refer to as the three major submarkets. First is non-perishable food, including consumables, canned food, pet food, health and beauty products, and pharmaceuticals. This is a large business that is still growing at a very healthy pace, and we are very satisfied with this business.

Decades ago, most warehouse clubs entered the grocery business in this way. Therefore, it is still growing at a very healthy pace. We also have physical stores and online stores at Whole Foods Market, a pioneer and leader in the organic grocery industry, which is still maintaining good growth momentum. Last year, we also made a series of reforms in terms of profitability, and we are very pleased with the profitability trajectory we see.

To meet the diverse grocery needs of our customers, we must have large-scale physical stores. This is something we have been working on for several years. We have tested the "Fresh" model in several locations near Chicago and in Southern California. We still need a little more time to observe, but from the results, it seems that we have something resonating with customers.

We have also been spending time and effort here to make it more convenient for customers to choose between non-perishable food, curated food, and fresh food. I think over time, you can see this in the user experience of our applications and websites, and you can also see how we better utilize different business units and their logistics capabilities to allow people to place orders in one place but pick up from multiple locations, and pick up multiple types of grocery products in one place. We have been doing more of these things, and I think we will continue to do so in the future.

In the healthcare field, if we consider the work we have done in retail, expanding into pharmacy services is a very natural extension. It is something that customers have been asking us to do for many years. It is more complex than our other retail businesses. Therefore, we had to carefully consider whether to enter this business, but customers really need it. We believe that this experience can be improved, and we can be a significant part of changing this experience, so we continued to do it. I am very pleased with the development momentum of Amazon Pharmacy.

It is growing very fast, but more importantly, if you have used it and paid attention to the customer experience in the past 12 to 15 months, you will find that it has made significant improvements on its already decent foundation.People really enjoy this kind of experience. I think that the healthcare experience, especially in the United States, is a very frustrating experience and not a very good one.

One of the important reasons why One Medical attracts us is their application. Their application is very user-friendly. You can access all your healthcare data in one place.

You can chat with doctors. If you need to see a doctor, you can also have a video call. It's a very different experience. If you really need medication, you can have it delivered to you within one or two days through Amazon Pharmacy or other pharmacies we partner with. This experience is much better than what we are used to seeing.

We have launched a monthly $9 or annual $99 healthcare subscription service for Prime members, which is 50% cheaper than the regular price. The acceptance of this service is very high. So, it's still early, but we believe we have the opportunity to become an important part of changing this experience.

If we can help customers change the primary care experience, change the way they access medication, then over time, we may be able to help them solve many other problems. Whether it's health or diet, I think over time, we can provide assistance in many aspects.

Q: Following up on Amazon. You outlined the generative AI stack, which I think is very clear. So I'm curious, given the competitive landscape at the application layer, how will you market at the application layer? And if you can, Andy, maybe you can expand on the consumer-facing AI strategy. I know you launched Rufus today. Do you think this can significantly improve the conversion rate and overall consumer engagement of retail applications, or do you have any vision in this regard? Thank you.

**A: Let's talk about generative AI first. When we talk to customers, especially enterprise customers, many are still thinking about which layer of the three-layer stack I listed they want to operate in. We predict that most companies will operate in at least two of these layers. But I also think that many technically capable companies will operate in all three layers. They will build their own models and use our existing models to build applications.

In the early stages of generative AI, we also see an interesting phenomenon, which is the process of going from inserting a problem into a chatbot and getting an answer to turning it into an application with production quality and achieving the quality required for customer experience and reputation. It's a very iterative process and a challenging task. And then, you also need to make sure that the application runs with the latency and cost characteristics you need.

So, we see that customers need choices. They don't want just one model to rule the world. They want different models for different applications. They want to try all different scales of models because these models have different cost structures and latency characteristics. Therefore, Bedrock really resonates with customers. They know they want to change all these variables and experiment and try. They need to manage all these different transitions and changes to find the best approach for them, especially in the first few years when they are learning how to build successful generative AI applications, which is very important for them.Therefore, this is also part of the reason why we see such a strong response to Bedrock.

Similarly, when companies consider coding partners like Q, what attracts them is the ability to improve developers' efficiency by 30%-40%. In many cases, developers are the most scarce resource for companies, so this changes the game. They don't promote every piece of code from coding partners. But if it can help them quickly achieve 80% of their goals, that's a big deal.

What sets Q apart is that it is not just a coding partner, it is an expert from Amazon that can help you write, debug, and test code. It can also help you with transformations, finding out how to achieve functionality in multiple steps, and troubleshooting. If your application has a problem, Q can find it and help you solve it. Therefore, it also allows you to view all data repositories, whether it's an intranet, a wiki, or over 40 data connectors like Salesforce, Atlassian, Zendesk, and Slack. It enables you to have intelligent conversations, get answers, and take action. So, this is a very unique feature.

When companies consider how to improve developers' efficiency, they consider the different features of these coding partner options. So, we spent a lot of time on this. Our company is quite excited about it. It has caused quite a stir at reinvent. What you usually see is that these companies try out the different options they have and make decisions for their employee base. We see very good momentum in this regard.

Regarding how we consider Gen AI in our consumer business, we are building dozens of generative AI applications across the company. Each of our businesses has multiple generative AI applications. They are at different stages, with many already launched and some still in development.

So, if you only look at some of our consumer businesses, in terms of retail, we have built generative AI applications that allow customers to view summaries of customer reviews, so they don't have to read hundreds or thousands of reviews to understand what people like or dislike about a product. We have launched a generative AI application that allows customers to quickly predict the fit of different clothing items. We have also built generative applications in fulfillment centers that can predict how much inventory we need at each specific fulfillment center.

Therefore, the launch of Rufus today is just another step, but we believe it is significant in becoming an AI-driven generative shopping assistant. It is trained on our extensive product catalog, our community Q&A, customer reviews, and the broader web, but it allows customers to discover products in a completely different way than they do on e-commerce websites. So, if you need buying advice, like how to choose a pair of headphones; if you are making a purposeful purchase, like what golf gear to buy for cold weather; or if you want to compare, like the difference between lip gloss and lip balm;Or maybe you want to recommend the best Valentine's Day gift; or you're browsing a detailed page with rich product information, but you don't want to go through the whole page. You want to ask if this tennis racket is suitable for beginners? You can input all these questions and get good answers, and then it seamlessly integrates into the Amazon experience, making customers accustomed to and enjoying taking action.

Therefore, I think this is just the next iteration. It will greatly change our shopping experience and the way customers discover. I can introduce these consumer businesses one by one. Our advertising business is building such features: people can submit a picture and the ad copy will be written, and vice versa.

Think about Alexa, we are building a very large-scale language model that will make Alexa more productive and helpful to customers. Each of our consumer businesses has a large number of generative AI applications, either already built and delivered, or in the process of being built. I don't think this situation will change for many years. We have a lot of ideas.

Risk disclosure and statement of this article: Dolphin Research Disclaimer and General Disclosure